• Difficult but very interesting course. Learned a lot. Too much benchmark comparisons.

• Very time consuming, but worthwhile if you want to get your hands dirty in OS.

  Should provide more consultation for project design issues (can learn more about advantages and disadvantages).
  I thought we had provided plenty of consultation.
  Here's what one of the tutors comments.

• Is it possible to `integrate' the project and lectures a little more - clearly not to the point of spoon feeding - but so that lectures such as the 64-bit page tables coincide with that particular phase of the project?
  I appreciate what you are saying. It's difficult to do without making the whole structure of the lectures fairly random. For example, the PT stuff used to be at the very end and I moved it forward as much as I thought I could (but still too late for the projects). There were a few other things which were out of logical order in hindsight.
  I'll keep working on it.

• Lots of work, but worthwhile.

• Please move the MIPS box on the table, pressing reset is hard.
  You're right, that was an oversight of mine when the lab was last set up. I have already given instructions that next year the two computers will be swapped around.

• Not enough focus on the low-level `hard' bits of an OS... in our project we are just coding to another API (L4's) In particular, thread scheduling, esp for multiprocessor systems. I mean practical considerations, like how and where you actually go lock things. Looking at the L4 source at the end is nice 'tho boring :-)
  Overall, good subject - this is a minor gripe.
  Thanks for these comments, I'll try to see what can be done about it. One thing I might be doing is to look closely as L4 source fragments throughout the course, rather than all in one shot. I'll work on it.

• Additional facilities would be nice:
  • library
  • coffee machine
  
  :-))
  As far as the "library" goes, all the recent stuff is available on-line. And for the older stuff I think it doesn't do too much harm if you guys occasionally walk the long way to the main library.
  As far as coffee goes, with the move to the new building we'd be keen to see the Union area between ME and our new building to cater for what our students want. However, we don't control the Union, so we'll have to see where we can go...

• The documentation could be a bit better. There are some undocumented "features" and in other places the documentation is misleading. At times the lectures also get a bit sidetracked.
  Tell me what's wrong with the dox and I'll fix them.
  As far as lectures go, do you really want me to go full power all the time?

• Stable/more bug-free kernel
  Easier said than done. The kernel used to be very stable but seemed to have developed a bit-rot problem :-)
  Looks like a simple recompilation exposed a bug that has always been there. No protection against that (and don't think anyone is developing OSes in a more stable environment).
  Still, this is the worst session so far for kernel problems, and it is a worry. The good news is that pretty much all problems could be traced back to a single bug which is now fixed. And there are now plenty people who know the kernel well, so chances are that the same sort of problems won't happen again.

• Too much lecture time on L4 system calls. We have the reference manual and can ask the lab demonstrators. For any other problems, I'd prefer we spent the lecture time on more interesting issues.
  Fair point. I think the dox are now good enough so I can indeed reduce the amount of time spent on the L4 API.

Final word